Phoenix: social media listening by and for peacebuilders and mediators

A few days ago, we launched Phoenix, an open source, free platform designed to enable peacebuilders and mediators to conduct ethical and participatory social media listening in order to inform conflict transformation work.

The tl;dr is that you can now use it in one of two ways:

• Peacebuilders and mediators can request access to the platform hosted for free by Build Up by filling out this application form.
• Developers can access the code in GitLab, and deploy it. Phoenix code is licensed under AGLP-3.0.

But if you want to hear more about what Phoenix can do and why we think it’s an important addition to the peacebuilding and mediation field, then read on…

What is Phoenix?

The Phoenix platform is user-friendly, and allows peacebuilders and mediators with no technical knowledge of social media listening to:

1. Gather posts and comments from social media platforms, based on search parameters they define;
2. Classify social media accounts as well as the text of the posts and comments, including with simple AI-assisted tools;
3. Explore the data they have collected from social media platforms, by searching it and building an interactive dashboard with graphs, tables and filters.

If you want more detail on how each of the gather, classify and explore modules work (including their limitations), it’s all documented in the Phoenix manual.

Why do peacebuilders and mediators need to conduct social media listening?

Patterns of digital content consumption and interaction are intertwined with polarization, anti-democratic practices, dehumanization, and violence. This problem is widespread and indiscriminate, impacting media users directly with broken relationships, harassment, and manipulation as well as indirectly with eroded institutions, tension, and violence, regardless of media use.

There is an abundance of experimental research showing correlation and cases demonstrating negative impact, but the time-consuming and intricate nature of the problem makes causal evidence, and solutions, illusive. It is a complex problem that paralyzes conflict responders and community peacebuilders, who are the ones best equipped to understand the conflictual dynamics found in a specific context but do not have the resources to focus on it. This leads to the crafting of uninformed interventions that risk causing more harm than good in the online dialogue fabric.

Understanding what digital harms stem from what happens on social media, and how they affect conflict, both online and offline, is critical for desining evidence-based dialogue and mediation initiatives today. Peacebuilders and mediators can benefit from the kind social media listening that is possible with Phoenix, because it provides a fine-grained understanding of what divides are playing out on digital media, how they are spreading, who is involved and their real-world impacts.

Our vision is for peacebuilders and mediators to be able to conduct social media listening without having to hire intermediary researchers or companies. Since 2020, Phoenix has been deployed in over 20 countries to support mediation teams, INGOs, UN agencies, and over 20 civil society organizations — in Arabic, Kurdish, Swahili, French and English. In 2023, Phoenix was selected as a featured solution by The Digital Public Goods Alliance and the United Nations Development Programme and showcased in the 2023 Nobel Prize Summit.

How can peacebuilders and mediators conduct social media listening?

We think social media listening can be used to answer many different kinds of questions relevant to peacebuilders and mediators. But overall, there are a few questions that come up for us over and over:

1. You can use social media listening to do a quick snapshot analysis of one piece of misinformation, understanding what it is, where it comes from and who it is impacting. This can be important when we are in a sensitive situation — elections, peace process — that can be derailed by misinformation.
2. You can use social media listening to track trends in narratives about a conflict dynamic over a period of time. We can look to understand what is being said on social media about a topic or topics (e.g. a group of people or an event or a political issue). Or we can look to understand what a group of people (e.g. youth groups or journalists) share on social media, and how that might connect to issues relevant to a topic. This can be important as a contribution to context analysis, in designing all manner of peacebuilding and mediation initiatives
3. You can use social media listening to investigate a problematic behaviour on social media (e.g. intimidation, hatespeech, harassment). This is useful when we are designing digital peacebuilding programs or want to introduce in a mediation process a clause relating to what conflict parties do on social media.

Often it’s some combination of the above, such as a problematic behaviour that happens when certain groups of people talk about certain topics. At Build Up, when we use Phoenix to conduct social media listening, we synthesise our rationale for doing a specific piece of analysis in what we call “problem statements”, which might be something like:

Conduct a mapping of online hate speech occurring at the local level in Coastal and Northern Kenya, understand who is using the language and to what end.

Understand how Yemenis are expressing ethnic and religious identities on social media.

Explore what narratives and behaviours are used to intimidate immigrants and people of colour on Instagram in the Netherlands.

Help us keep Phoenix free!

Phoenix was developed by Build Up and datavaluepeople, with the support of the Swiss Federal Department of Foreign Affairs. Phoenix is a digital public good: the code is open source and we aim to host a free platform for approved users.

If you are interested in making a donation to help us maintain the code and keep the platform free, you can take a look at what we need funding and make a donation on this page, or contact us at phoenix@howtobuildup.org